Summary RP3 Hierarchical statistical modeling and causal inference approaches to elucidate exposure pathways underlying health disparities The health disparity between the Native American population and the US general population arises from the complex interplay between multiple socio-demographic, behavior, lifestyle and genetic susceptibility factors. Environmental contaminants are increasingly acknowledged to play an important part in explaining health disparity through their combined or interaction effects with other factors. Proximities of Native American communities to abandoned uranium mines (AUM) have been of particular health concern. These chronic exposures to AUM waste related metal mixtures pose higher risk for developing chronic and fatal diseases including hypertension, diabetes, kidney disease, and types of cancer in Native American populations compared to the US population. The hypothesis of this project is that the three Native American tribal communities included in this study (Navajo Nation, Crow, and Cheyenne River Sioux) encounter great risk of exposures to environmental hazards (mine waste related metal mixture exposures, unregulated water resources, and illegal dumping, etc.). These hazardous exposures along with socioeconomic status, psychosocial stress, behavior/lifestyle factors influence multiple biological pathways to produce health disparities in Native American communities. The complex set of exposure variables including dietary nutrients, physical activity, infectious agents, air pollutants and metal exposures at both the individual and community levels are acknowledged as contributors to health disparities, however, their relative contributions of the potential causal factors have not been well studied. The objective of this project is to employ data-driven and modeling approaches to understand the relative contribution of different environmental, behavior, and socioeconomic determinants of the health disparities between the native population and the US national population. We will use innovative modeling approaches such as decomposition analyses and structural causal models to estimate the effects of risk factors at the individual and community level on the health disparities. In Aim 1, we will collect data and summarize the frequency distributions for major chronic and fatal diseases in the Native American communities. In Aim 2, we will employ novel hierarchical modeling approaches to estimate the relative contribution of different risk factors at the individual level and community level to the health disparities. In Aim 3, we will implement frontier causal pathway analyses to illustrate the intermediate mechanisms explaining the health disparity. Aim 4 is to examine the complex correlation structure among multi-dimensional exposures, intermediate biological responses, and health endpoints using frontier statistical approaches. We expect this project will identify major contributing factors that explain a large proportion of the health disparity, and in addition elucidate the intermediate causal pathway that the effects are transmitted to the health disparity endpoints. These findings have the potential to inform policymaking on the cost-effective resource allocation to maximally reduce disparity and improve community health.